Geoff Webb
Vice President Solutions, Product, and Marketing Strategy isolved Follow

On today’s episode of the RecruitingDaily Podcast, William Tincup speaks to Geoff from isolved about Ai and chatbots and whether or not they should be trusted to onboard new hires.

Some Conversation Highlights:

Actual real things, yeah, that actually worked, which has been nice, because it’s given me an opportunity to dig in. Mostly it gives me an opportunity to sit down with people that actually really, really understand it and actually get them to explain it to me over and over again so that I can follow it. But NLP, natural language processing, is one of the things that you can do with AI, where it’s looking for ways to interact with human beings in the way that we are comfortable. Human beings are language machines. That’s what we do. We make words. And when we can’t think of a word for something, we make up a new word for it. We’re amazing communicators. It’s so foundational to even the human condition, this desire to communicate.

What AI can do is, when appropriately built and properly trained and correctly delivered, and there are so many caveats in there, but when you do it right, it can enable people to communicate with other technology devices in a way that feels much more natural. Literally the natural language processing part, the natural kind of communication, the cadence of communication, the style. And good natural language processing tools that are often built into chatbots will learn. They can learn not just what questions are being asked, but even how individual users are asking those questions, so that they can learn that when they use this word, they really mean this.

Tune in for the full conversation.

Listening time: 26 minutes

Codesignal Diverse Companys Outperform

 

Enjoy the podcast?

Thanks for tuning in to this episode of The RecruitingDaily Podcast with William Tincup. Be sure to subscribe through your favorite platform.

Announcer: 00:00 This is RecruitingDaily’s Recruiting Live Podcast, where we look at the strategies behind the world’s best talent acquisition teams. We talk recruiting, sourcing and talent acquisition. Each week, we take one overcomplicated topic and break it down so that your three-year-old can understand it. Makes sense? Are you ready to take your game to the next level? You’re at the right spot. You’re now entering the mind of a hustler. Here’s your host, William Tincup.

William Tincup: 00:34 Ladies and gentlemen, this is William Tincup and you are listening to the RecruitingDaily Podcast. Today we have Geoff on, from isolved, and our topic is, should AI and chatbots be trusted to onboard new hires? I’m excited about this topic, so let’s just jump right into it. Geoff, would you do us a favor and introduce both yourself and isolved?

Geoff Webb: 00:56 Yeah, absolutely. I would love to. So isolved is a business. We are focused on enabling companies to just deliver an exceptional employee experience. We provide that through both a cloud-based HCM platform that kind of runs pretty much everything, which is going to be good for us to talk about today, from onboarding all the way through, and even hiring, all the way through to benefits and workforce management and the rest of those things that are important, and collaboration tools and so on. And also rolling in a lot of expertise that’s sort of on demand. So HR, augmentation services too.

01:29 So we do a lot of that. Maybe about five million people in the US touched by our technology and our services, for about 150,000 companies around the US. So we’ve got some good reach, which gives us a good perspective. For myself, I lead the product and solution marketing and strategy team. So I spend a lot of time looking at emerging technologies, looking at the shape of things as they emerge in the overall HCM, HR world and the intersection of that with technology, and then making sure that we as a business stay focused on the right things, pointed in the right direction. So it’s a really interesting job. I love where we’re at right now, and this topic is absolutely the place to be having this conversation and this is a great topic to be talking about.

William Tincup: 02:15 Well, as a recovering marketer, one of the things I would tell the audience is what I love about isolved is it’s wonderfully marketed. I mean, it’s beautiful brand. Everything is just… It doesn’t look like an HR tech company, which is the highest compliment I could possibly give isolved, because most of the marketing that’s done in HR tech or rec tech or whatever is really boring and really ugly, quite frankly. So isolved has transcended that. It’s just a beautiful brand, well put together, everything. So whomever, I’m sure there’s a bunch of people that had a lot to do with that, but it’s just wonderfully done.

Geoff Webb: 02:59 No, thank you.

William Tincup: 03:00 Yeah. Let’s start with the difference between AI and chatbots and the way that you think of it for the audience. What is the difference, in their minds, or what is a difference in your mind between what is AI as it relates to this topic, and chatbots?

Geoff Webb: 03:22 Yeah, of course. Well, first of all, I would say that chatbots, many of them will use some version of AI, of artificial intelligence, to do a couple of things. One is to respond more intelligently and the other is to learn as they respond so that they can continue to track and stay ahead of what people are asking them. So a lot of chatbots use AI, but artificial intelligence overall, despite honestly being horrifically over-marketed and over-hyped over the past few years, and thrown into every really cheesy horror film and science fiction film you could possibly imagine, actually is very, very pervasive in a lot of technologies already, is already delivering a huge amount of value, often in places you don’t expect and in ways that you don’t really think about.

04:11 And so it’s such a ubiquitous technology, and set of technologies, and it does some things really, really well. And chatbots is certainly an example of that, but its reach is so broad. It’s going to be challenging, I think, in a few years to find technologies that don’t, in some way, lean on some kind of AI engine that’s delivering some extra capabilities.

William Tincup: 04:34 Do we even get into the difference between NLP or machine learning and AI? Because I love the… And I believe you’re spot on with the over-hyped part. And I think it’s both sides. It’s vendors. We’ll go to a conference like HR Tech, and vendors, you go from booth to booth and everyone says AI. Okay, so we got that. Then you’ve got on the practitioner side across the spectrum that you’ll deal with, I’m not sure they really know what AI is. So I’m actually not sure that the vendors outside of people that are in product and tech, I’m not sure that they know the difference between any of those things.

05:20 I did this bit at Greenhouse Open in May where I said, “Listen, I just don’t believe that practitioners actually know the difference between blockchain, Bitcoin, NLP, AI.” I just listed all these topics, right? And so the person I was with, they’re like, “No, they actually do, William. You’re just being really pessimistic.” I’m like, “No, I really believe that.” And I said, “Just pick somebody in the crowd.” So they said, “Okay, that guy.” So I literally walked up to him and said, “Hey, just doing kind of a straw poll here. What’s the difference between NLP and machine learning and AI? Just basic stuff.” They said, “I have no idea what you’re talking about.” No, no, I grin. A survey of one does not constitute…

Geoff Webb: 06:06 Small sample. Not very helpful. Not helpful. Yeah.

William Tincup: 06:12 What’s your take on people’s literacy when it comes to these things?

Geoff Webb: 06:18 Also, I think this is a good question. I’ve been lucky, actually. I’ve worked for a few AI technology companies that were actually doing actually real things with AI.

William Tincup: 06:26 Actual AI. Yes.

Geoff Webb: 06:28 Actual real things, yeah, that actually worked, which has been nice, because it’s given me an opportunity to dig in. Mostly it gives me an opportunity to sit down with people that actually really, really understand it and actually get them to explain it to me over and over again so that I can follow it. But NLP, natural language processing, is one of the things that you can do with AI, where it’s looking for ways to interact with human beings in the way that we are comfortable. Human beings are language machines. That’s what we do. We make words. And when we can’t think of a word for something, we make up a new word for it. We’re amazing communicators. It’s so foundational to even the human condition, this desire to communicate.

07:09 And what AI can do is, when appropriately built and properly trained and correctly delivered, and there are so many caveats in there, but when you do it right, it can enable people to communicate with other technology devices in a way that feels much more natural. Literally the natural language processing part, the natural kind of communication, the cadence of communication, the style. And good natural language processing tools that are often built into chatbots will learn. They can learn not just what questions are being asked, but even how individual users are asking those questions, so that they can learn that when they use this word, they really mean this.

07:51 And so there’s great potential yet to be realized, I think, in that space of natural language processing and human interaction with AI that can unshackle us from a lot of the artificial communication media that we’ve been required to use, things like keyboards and mice that we learned to use, but they weren’t natural to a human being. Language is so natural. So I think AI, one of the great things it can do, AI can enable us to communicate with technology and with each other in ways that is much more natural to the way we think and the way we typically communicate. But again, you’ve got more broad AI, there’s a lot of other things going on there that we can certainly explore.

William Tincup: 08:37 For those that are listening, the word “potential,” when you’re in your twenties and thirties, that’s a compliment. When you’re 40 plus, “potential” is actually not a compliment. So keep that in the back of your mind. Let’s talk a little bit about onboarding and your take on how… First of all, we’re talking about should, so let’s just answer that question. Should these things be used in onboarding of new employees?

Geoff Webb: 09:12 I think the short answer to that is there is a lot of good reasons to use AI technology in the onboarding process and in the various steps they’re in, for things that we can definitely talk about. There are things that AI does exceptionally well, again, when built properly and trained properly, that add huge amounts of value to the experience of onboarding for everybody that’s involved. And I think that’s going to be the key to saying whether this is a good idea or not, is delivering a better experience for the employee during that onboarding process, and even before onboarding, to be honest.

09:49 And I think also for the employer, all the people that are… the recruiters, the hiring managers, the folks that are administrating these things, giving a better experience to them too. I think there’s a lot of friction in a lot of processes that have grown over time that AI can help us cut through in a bunch of different ways. So, should we? Yes, absolutely. And are we? Yeah, companies absolutely are. But again, you’ve got to approach this as it’s an enabling technology. It isn’t somehow a magic wand. It isn’t going to fix everything. It’s going to enable you to do a better job at certain things.

William Tincup: 10:27 I agree, obviously, and I see it as connective tissue between disparate candidate experience and employee experience. Onboarding has historically been the baton pass between recruiting and HR. And if you ever watch a four by four relay, batons get dropped. That happens quite often. And I think this is true of onboarding, historically at least, that we romance candidates. They say yes, offer letter accepted, start date, all that type stuff. And then all of a sudden, the romance is over.

11:10 Again, I think there’s ways to extend that experience. Because you touched on something I think is also fascinating, that the audience should hear about is, a lot of this movement of pre-boarding, where after the offer letter has been signed, but before the start date, there’s some things that can be done that I believe AI and chatbots can have a wonderful hand in creating. Again, there’s disparate experiences because they’re dealing with different teams of people. So what’s your take on pre-boarding?

Geoff Webb: 11:47 Oh, yeah. Yeah, yeah, yeah. I mean, we’ve all been through this, I’m sure, of that sort of initial… the excitement. And the thing is, the process of going through interviewing and working your way through all of that, and then that moment of triumphal… the accolades you receive of, “Yes, we want you. You’re the person. You’re going to join this company,” and you have this amazing sense of, “I’ve achieved something. I’ve finally… I’ve reached the summit of this particular opportunity.” And then all of it goes dead and just falls completely flat, because then it’s three weeks later and you haven’t heard anything. And that’s a miserable experience for the employee and it sets such the wrong tone for everything else.

12:27 We’ve done studies of employees in the US, and we asked about the impact of that, the whole onboarding process. And it’s absolutely devastating if you don’t get it right. We saw, I think, one statistic, there was something like it was well over half of the employees that we asked had said, you know what, they actually considered going somewhere else because the onboarding was so bad, and the pre-boarding is absolutely a critical part of that. Because that’s the moment of truth. That’s the moment from, as you say, we’ve gone from that sort of initial… the interviewing process and that moment of almost the honeymoon is here, and then when you switch and suddenly it goes quiet and you don’t hear anything or you can’t get answers or you’re not sure what to do next and you don’t know who to turn to, it becomes a very unpleasant and unnerving sense of, “Wait, did they forget about me? Am I really not that important?”

13:20 And I think, to go back to the original topic, that things like chatbots are great tools, AI generally, great tool for handling the basic stuff that you might ask. Like, “Well, okay. Oh, I forgot to ask, what are the company holidays? What’s the organizational structure I’ll be working in? If I need to move my start date, can I do that? What are the benefits over here? I forgot to ask about this benefit, or can you remind me of this?” So many transactional points that I think that aggregate into a good experience if done well also aggregate into a really bad experience if ignored or not responded. And they’re basic and you can fix that with simply being available. And I think one of the availability tools can be chatbots and AI overall, to help navigate that conversation to a much, much better outcome.

William Tincup: 14:06 And it’s the questions. And again, there’s a library of questions that anyone would ask from candidate to employ that transition. And again, having AI and chatbots to then be able to answer those questions, because people are working from home, working all around the world, different time zones. So it’s not as easy as walking down to HR or whatever and asking a question, which of course is hard enough. To be able to ask a chatbot a question and say, “Hey, quick question about,” I loved the one you said, “Can I change my start date?” That’s a great question.

14:46 But I also think that there’s a reactive part that chatbots and AI can play a part in, in that, “Okay, you’re a candidate and now an employee, freshly [inaudible 00:14:59] employee. You have questions. Okay, there’s a bank of answers over here. Or if there’s not, we’ll find somebody that can answer that question.” But also I think there’s probably an element of preemptive, where, “Here’s some of the questions that commonly come up and do you have these questions or if you don’t have these questions, do you have any other questions that we can handle so that you have a wonderful experience?”

Geoff Webb: 15:24 Yeah, I think that’s exactly right. One of the great values that AI has is not just its capacity to navigate the transaction effectively, but to learn and to respond preemptively, as you say. So the nice thing about AI is it’s really good… Generally, AIs are good at looking at and understanding large amounts of information around trends and common events and looking for patterns. And they can very quickly identify, “This is a common set of questions I get asked. Why don’t I help provide the information up front?” So you don’t feel like you’re having to be pulling teeth, and then we can deal with the exceptions beyond that with the natural language processing.

16:09 And I think the other thing that’s really interesting that comes out of that is because you’re able to start to identify what questions are coming up regularly, as a recruitment team, you can start to identify, “Well, actually we probably should spend more time on this because everybody asks about it. So maybe we should fix that up front and make sure everyone understands that. Or maybe we should make sure there’s some materials we can send out in advance to preempt it so they don’t feel like they have to ask about it.” So I think there’s the combination of AI as a responsive technology, but also AI as an analytic technology and even a predictive technology to help you shape a much better experience in a whole bunch of places. And again, as we talked about, that onboarding is a critical place to get it right.

William Tincup: 16:54 It’s funny, because onboarding used to be a binder or an hour or a day or a week. It was this thing, right?

Geoff Webb: 17:04 Yeah.

William Tincup: 17:04 You’d go in a room and hope that things worked out. But it’s so different now. The expectations, I guess, is really the expectations. And I won’t even put this in generational. I think the expectations have just shifted to consumer experience. I want to have a consumer experience at work. I want to have that. I want to be able to go to any website, there’s a chatbot there that I can ask it questions. And my sons are both Gen Z and they’d rather talk to chatbots, which is a different problem, or a different opportunity, if you will, because they’d rather not interact with humans and just interact with chatbots.

17:51 But I wanted to ask you, because we’ve both used the phrase, “If done well.” So for the audience, how do we set that up? What are examples of if done well, especially as it relates to onboarding and chatbots and AI. Let’s talk about if done well. We don’t have to talk about the catastrophes and not done well, but if done well, what does that look like? And maybe if you have some examples, that’d be great too.

Geoff Webb: 18:18 Yeah. Well, I think some of the examples we’ve already touched on, and yeah, you’re right, there are plenty of horrific examples of AI in the world of conversation that are probably best left on the ground and not touched. I think it’s a case of enabling that conversation to be very natural and to be very… The worst is when you’re trying to guess, “How do I ask this question so I can get an answer? What word do I have to use so that I can get the answer on benefits? Do I have to ask it like this, or I have to put a capital B?” Because it’s crazy stuff that you see when we’re trying to interact with technology that’s poorly put together.

18:55 I think what works really well is when you’ve got the technology in a place where it can respond very naturally to a whole range of different kinds of questions without requiring the human element to start to guess how it wants to operate. I don’t want to act like a robot. I’m happy to talk like a human being. I think the second thing is that you have things like chatbots that are able to guide. So in other words, if you ask a question that’s like, “I don’t know what that is,” have it help, have it guide, “Okay. Are you asking about this? Maybe that’s what you’re talking about. Is it something like this?”

William Tincup: 19:34 I love that. First of all, I love that because it’s the probative nature of it, and it’s like someone says, going back to the example, “May I switch my start date?” It’s like, “Okay. Yes. So let’s get some more specifics around that. Are you thinking about a day? Are you thinking about you need to push it back a week, et cetera.” Again, technology can do that.

Geoff Webb: 20:00 Absolutely.

William Tincup: 20:01 We can do it as humans, but technology can do that and let humans do other things.

Geoff Webb: 20:06 Yeah. And I think there’s a few things there. You’re absolutely right. First of all, the technology is quite capable of doing that, if you invest appropriately in building it that way. I think the other thing I would say is that one of the really nice thing is if you can continue that capacity to have the conversations through all the steps of the employee’s life with the business. So in other words, chatbot, absolutely outstanding, great tool for those onboarding questions, but continuing that capacity to have that very natural interaction through all the other steps and all the other things that HR and the rest of the business will do.

20:44 So in other words, to have that, and I hesitate to use the word, relationship with the AI chatbot, but you know what I mean. It’s a sense of, I can go to it and ask it almost anything and have it connected to all the different elements. That’s really powerful, because now I have a place that I can always go and say, “I don’t remember. How do I do this? Can you tell me who does that? Or how would I get access to this, or what’s the best way to do that? So that from pre-boarding to onboarding, to every single day, every interaction, there’s a consistent way for me to interact with the fabric of the human resources infrastructure in a way that is very natural and that elicits interaction.

21:25 And that’s so powerful, because now you’re switching from, “Oh no, I don’t want to get… I’ll put this off. I want to ask that question, but I won’t ask it now because I’ve got better things to do,” to something that is much more natural and immediate and feels like, “Actually, there’s something there that’s there to support me.” That changes the rules and it really reshuffles a deck when it comes to thinking about how I experience working with this business.

William Tincup: 21:49 It’s funny, because as you were talking about it, one of the things that I think about is it’s not just the employee, there’s a whole ecosystem around the employee, a partner, spouse, wife, husband, whatever, children. There’s a bunch of other questions that that person might have that aren’t even their questions, technically. And so again, they might not raise their hand for fear of maybe they don’t want to come across and ask that question that’s difficult or whatever. So I love the idea of, “Okay, it’s just chatbot. You’re good. Not going to get offended. You’re good.”

Geoff Webb: 22:28 Exactly. I’ve seen this in a bunch of other technology areas too. Healthcare technology, those kind of areas. AI’s do not have… They don’t judge. An AI is a judgment free zone. You can ask it the dumbest question ever that you’d be totally embarrassed to ask and it will be happy to answer it and will not judge you for it and it will move on with [inaudible 00:22:48].

William Tincup: 22:48 It’s like it’s a dog. Dogs are always happy to see you. It doesn’t matter.

Geoff Webb: 22:52 Stupid.

William Tincup: 22:54 Come in the door, dogs are like… I have two pit bulls. Every time I walk in the house, it’s like they haven’t seen me in weeks. I love it. The last thing I wanted to get your take on is we talk about the build process, because I know the audience will probably have these questions in their mind. If something’s broken, you can’t replicate what’s broken with AI and chatbots and expect a different experience. So what do they need to do to kind of audit and create something great, and then replicate it in AI, in chatbots?

Geoff Webb: 23:32 Yeah, you’re right. Actually, in fact, it’s worse than that. Not only will it not fix it, it’ll almost certainly make it worse.

William Tincup: 23:37 Right. 7%.

Geoff Webb: 23:40 So I think that the first thing is, you’re right. You’re a hundred percent correct. You have to start with a really cold, honest, sober look at the process today, and that means you’ve got to go back and talk to people that have been onboarded. And I’ve done this in companies, talked to people and had that conversation, “How was the onboarding experience? What was it like and what was the worst part?” And man, you hear some different stories. And so, you have to do it. You have to figure out what isn’t working and then you have to start to deploy… and I would say incremental is always a good thing here… incremental improvements that you can measure and then focus on what’s the next worst thing to go fix.

24:21 Trying to do a moonshot, solve everything in one go, it can work, but it can also be an awfully draining process. I would say incremental improvement. Look for trends. Ask people. Listen to the answers. Don’t come in with a, “No, that’s not what we’re like.” If everybody’s telling you that this is their experience, then that’s probably their experience.

William Tincup: 24:47 Yeah. I think there’s an element of vulnerability here that we probably haven’t talked about, but probably a separate podcast in and of itself, is that you’ve got to ask them pretty much at every step of the way, “How are we doing? What would you change? What could we do to make this process…” And again, luxury hotels, as all luxury brands, they do a marvelous job at this. It’s almost like a finger on the pulse. They’re constantly getting feedback so that they can improve their process.

Geoff Webb: 25:17 And interestingly, that’s one of the other areas that AI can help. AI actually can be used very effectively as a steering mechanism for the experience. “This is what’s happening. This is a common set of behaviors. It takes this amount of time to do this trend. This is trending in this direction.” There are lots of things you can do with AI on the analytics side to just give you some steering towards what can be improved and what should be improved, and what’s happening when you try to improve it.

William Tincup: 25:44 I love it. Geoff, this has been absolutely fantastic. Thank you so much for your wisdom and your time.

Geoff Webb: 25:51 Oh no, absolutely. Thanks so much for having me. I’ve really enjoyed it.

William Tincup: 25:54 Absolutely. Thanks for everyone listening to the RecruitingDaily Podcast. Until next time.

Announcer: 25:59 You’ve been listening to the Recruiting Live Podcast by RecruitingDaily. Check out the latest industry podcasts, webinars, articles and news at…

The RecruitingDaily Podcast

Authors
William Tincup

William is the President & Editor-at-Large of RecruitingDaily. At the intersection of HR and technology, he’s a writer, speaker, advisor, consultant, investor, storyteller & teacher. He's been writing about HR and Recruiting related issues for longer than he cares to disclose. William serves on the Board of Advisors / Board of Directors for 20+ HR technology startups. William is a graduate of the University of Alabama at Birmingham with a BA in Art History. He also earned an MA in American Indian Studies from the University of Arizona and an MBA from Case Western Reserve University.


Discussion

Please log in to post comments.

Login